panel.matrices throws error if panel argument is a tidyverse tibble

[skranz]

Just a short notice. The panel.matrices function seems to throw an error if the passed panel object is a tibble (the tidyverse equivalent of a data frame). Here is a simple example with resulting error message:

data('california_prop99')
library(dplyr)
setup = panel.matrices(as_tibble(california_prop99))

Error in panel.matrices(as_tibble(california_prop99)) : 
  There is no variation in treatment status.

Since tidyverse is very popular and certain dplyr operations tend to convert data frames into tibbles, this error may be encountered by some users. A simple fix would probably be to start the panel.matrices function with a call panel = as.data.frame(panel). At least the following code works fine:

data('california_prop99')
library(dplyr)
setup = panel.matrices(as.data.frame(as_tibble(california_prop99)))

[yang-yuchuan]

THANKS A LOT!!! Your comment really helps me!

[MatthieuStigler]

The problem comes from unique(panel[, treatment]), which has a different implicit behavior with respect to the implicit drop argument

• if data.frame, selecting one column will convert to vector (drop's behavior is context-specific)
• if tibble, selecting one column will keep as tibble (drop =FALSE always).

This is why length(unique(panel[, treatment])) returns a false positive warning, compare:

df <- data.frame(treatment =c(0,1))
tib <- tibble::tibble(treatment =c(0,1))

length(unique(df[, "treatment"]))
#> [1] 2
length(unique(tib[, "treatment"]))
#> [1] 1

^{Created on 2022-09-12 with reprex v2.0.2}

[siminl]

Thank you so much! This is so helpful!!